linux-kernel - Re: [PATCH 1/2] mm/mprotect: Call arch_validate_prot under mmap

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Thu, 08 Oct 2020 21:34:26 +1100
From:   Michael Ellerman <mpe@...erman.id.au>
To:     Jann Horn <jannh@...gle.com>,
        Christoph Hellwig <hch@...radead.org>,
        Benjamin Herrenschmidt <benh@...nel.crashing.org>,
        Paul Mackerras <paulus@...ba.org>,
        linuxppc-dev@...ts.ozlabs.org
Cc:     "David S. Miller" <davem@...emloft.net>,
        sparclinux@...r.kernel.org,
        Andrew Morton <akpm@...ux-foundation.org>,
        Linux-MM <linux-mm@...ck.org>,
        kernel list <linux-kernel@...r.kernel.org>,
        Khalid Aziz <khalid.aziz@...cle.com>,
        Anthony Yznaga <anthony.yznaga@...cle.com>,
        Catalin Marinas <catalin.marinas@....com>,
        Will Deacon <will@...nel.org>,
        Linux ARM <linux-arm-kernel@...ts.infradead.org>,
        Dave Kleikamp <shaggy@...ux.vnet.ibm.com>
Subject: Re: [PATCH 1/2] mm/mprotect: Call arch_validate_prot under mmap_lock and with length

Jann Horn <jannh@...gle.com> writes:
> On Wed, Oct 7, 2020 at 2:35 PM Christoph Hellwig <hch@...radead.org> wrote:
>> On Wed, Oct 07, 2020 at 09:39:31AM +0200, Jann Horn wrote:
>> > diff --git a/arch/powerpc/kernel/syscalls.c b/arch/powerpc/kernel/syscalls.c
>> > index 078608ec2e92..b1fabb97d138 100644
>> > --- a/arch/powerpc/kernel/syscalls.c
>> > +++ b/arch/powerpc/kernel/syscalls.c
>> > @@ -43,7 +43,7 @@ static inline long do_mmap2(unsigned long addr, size_t len,
>> >  {
>> >       long ret = -EINVAL;
>> >
>> > -     if (!arch_validate_prot(prot, addr))
>> > +     if (!arch_validate_prot(prot, addr, len))
>>
>> This call isn't under mmap lock.  I also find it rather weird as the
>> generic code only calls arch_validate_prot from mprotect, only powerpc
>> also calls it from mmap.
>>
>> This seems to go back to commit ef3d3246a0d0
>> ("powerpc/mm: Add Strong Access Ordering support")
>
> I'm _guessing_ the idea in the generic case might be that mmap()
> doesn't check unknown bits in the protection flags, and therefore
> maybe people wanted to avoid adding new error cases that could be
> caused by random high bits being set?

I suspect it's just that when we added it we updated our do_mmap2() and
didn't touch the generic version because we didn't need to. ie. it's not
intentional it's just a buglet.

I think this is the original submission:

  https://lore.kernel.org/linuxppc-dev/20080610220055.10257.84465.sendpatchset@norville.austin.ibm.com/

Which only calls arch_validate_prot() from mprotect and the powerpc
code, and there was no discussion about adding it elsewhere.

> So while the mprotect() case
> checks the flags and refuses unknown values, the mmap() code just lets
> the architecture figure out which bits are actually valid to set (via
> arch_calc_vm_prot_bits()) and silently ignores the rest?
>
> And powerpc apparently decided that they do want to error out on bogus
> prot values passed to their version of mmap(), and in exchange, assume
> in arch_calc_vm_prot_bits() that the protection bits are valid?

I don't think we really decided that, it just happened by accident and
no one noticed/complained.

Seems userspace is pretty well behaved when it comes to passing prot
values to mmap().

> powerpc's arch_validate_prot() doesn't actually need the mmap lock, so
> I think this is fine-ish for now (as in, while the code is a bit
> unclean, I don't think I'm making it worse, and I don't think it's
> actually buggy). In theory, we could move the arch_validate_prot()
> call over into the mmap guts, where we're holding the lock, and gate
> it on the architecture or on some feature CONFIG that powerpc can
> activate in its Kconfig. But I'm not sure whether that'd be helping or
> making things worse, so when I sent this patch, I deliberately left
> the powerpc stuff as-is.

I think what you've done is fine, and anything more elaborate is not
worth the effort.

cheers